The joy of teaching AI

Teaching AI at LUISS and LUISS Business School has a useful side effect: it exposes the gaps in my own understanding.

An idea can feel clear when I am reading a paper or talking to someone who uses the same vocabulary. Explaining it to a room with different backgrounds takes more work. I have to decide what matters, find an example, and work out which details I can leave aside without making the explanation misleading.

That preparation is one of the parts I enjoy most. If I need a long list of technical terms to explain something, I usually need to spend more time with it myself.

Different questions

I teach both students and executives. They bring different concerns to the same subject. A discussion can move from how a model learns to what happens when its answer is wrong, or how a business would decide whether to use it.

Those questions help me connect the technical material to decisions people actually face. They also challenge assumptions that can go unnoticed when everyone in a conversation has similar experience.

I leave some classes with a better explanation than the one I arrived with. Sometimes I leave with a question I need to investigate. Both are useful.

Keeping the material honest

AI changes quickly, so preparing a class means revisiting examples and checking what still holds. I try to distinguish the ideas that last from the tools that happen to be getting attention that week.

That discipline carries into my work as an entrepreneur. A convincing demonstration is a start. Understanding what it depends on, where it fails, and who would use it takes more thought.

The pleasure of teaching comes from working through those questions together. I can share what I know, listen carefully, and come back better prepared for the next class.